better loop/branch target alignment
Bug #383952 reported by
Nikodemus Siivola
This bug affects 1 person
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| SBCL |
Incomplete
|
Low
|
Unassigned | ||
Bug Description
Currently on x86-64 SBCL aligns loops, but it would probably be better to align all backwards branch targets. ...but as has been noted, effects of alignment on modern processors seem to be getting harder to predict. Possibly a better strategy would be to try to be smarter about fitting loops inside cache lines, etc.
Revision history for this message
| Douglas Katzman (dougk) wrote | #1 |
| Changed in sbcl: | |
| status: | Confirmed → Incomplete |
To post a comment you must log in.
this is some combination of fixed and/or obsolete